题目：Some Results in Functional Time Series Analysis: from Structural Approximation to Simultaneous Inference

汇报人： 周舟

会议时间：2025年6月16日（周一）15:00

地点：综合楼644会议室

报告人简介：

周舟于2009年在芝加哥大学获得统计学博士学位。目前，他是多伦多大学统计科学系终身正教授。周教授的主要研究兴趣包括复杂时间序列分析，非参数和半参数推断，时频分析，变点分析，以及函数型和纵向数据分析。周教授于2021年获得了NSERC Discovery Accelerator奖，并于2023年获得了CRM-SSC奖。  

报告摘要：

Some recent results in functional time series analysis will be presented in this talk. First, a unified functional auto-regressive approximation result for a general class of locally stationary functional time series will be delivered which serves as a theoretical foundation for a range of inferential problems for non-stationary functional time series analysis, including optimal linear forecasting, covariance structure inference, and resampling. Second, we will discuss the problem of simultaneous inference for functional time series linear regression. In particular,  joint simultaneous confidence band construction for time series scalar-on-function linear regression will be investigated when the regression model is estimated by a roughness penalization approach.